Internationally distributed organic and inorganic oxygen isotopic reference materials have been calibrated by six laboratories carrying out more than 5300 measurements using a variety of high-temperature conversion techniques (HTC) in an evaluation sponsored by the International Union of Pure and Applied Chemistry (IUPAC). To aid in the calibration of these reference materials, which span more than 125‰, an artificially enriched reference water (δ¹⁸O of +78.91‰) and two barium sulfates (one depleted and one enriched in ¹⁸O) were prepared and calibrated relative to VSMOW2 and SLAP reference waters. These materials were used to calibrate the other isotopic reference materials in this study, which yielded:

The seemingly large estimated combined uncertainties arise from differences in instrumentation and methodology and difficulty in accounting for all measurement bias. They are composed of the 3-fold standard errors directly calculated from the measurements and provision for systematic errors discussed in this paper. A primary conclusion of this study is that nitrate samples analyzed for δ¹⁸O should be analyzed with internationally distributed isotopic nitrates, and likewise for sulfates and organics. Authors reporting relative differences of oxygen-isotope ratios (δ¹⁸O) of nitrates, sulfates, or organic material should explicitly state in their reports the δ¹⁸O values of two or more internationally distributed nitrates (USGS34, IAEA-NO-3, and USGS35), sulfates (IAEA-SO-5, IAEA-SO-6, and NBS 127), or organic material (IAEA-601 benzoic acid, IAEA-602 benzoic acid, and IAEA-600 caffeine), as appropriate to the material being analyzed, had these reference materials been analyzed with unknowns. This procedure ensures that readers will be able to normalize the δ¹⁸O values at a later time should it become necessary.

The high-temperature reduction technique for analyzing δ¹⁸O and δ²H is not as widely applicable as the well-established combustion technique for carbon and nitrogen stable isotope determination. To obtain the most reliable stable isotope data, materials should be treated in an identical fashion; within the same sequence of analyses, samples should be compared with working reference materials that are as similar in nature and in isotopic composition as feasible.